The present invention relates to semiconductor radiation emitter packages such as, for example, light emitting diode packages.
Semiconductor optical emitter components such as light emitting diode (LED) devices have become commonplace in a wide variety of consumer and industrial opto-electronic applications. Other types of semiconductor optical emitter components, including organic light emitting diodes (OLEDs), light emitting polymers (LEPs) and the like may also be packaged in discrete components suitable as substitutes for conventional inorganic LEDs in these applications.
Visible LED components of all colors are used alone or in small clusters as status indicators on such products as computer monitors, coffee makers, stereo receivers, CD players, VCRs and the like. Such indicators are also found in a diversity of systems such as instrument panels in aircraft, trains, ships, cars, trucks, minivans and sport utility vehicles, etc. Addressable arrays containing hundreds or thousands of visible LED components are found in moving-message displays such as those found in many airports and stock market trading centers and also as high brightness large-area outdoor TV screens found in many sports complexes and in some urban billboards.
Amber, red and red-orange emitting visible LEDs are used in arrays of up to 100 components in visual signaling systems such as vehicle Center High Mounted Stop Lamps (CHMSLs), brake lamps, exterior turn signals and hazard flashers, exterior signaling mirrors and for roadway construction hazard markers. Amber, red and blue-green emitting visible LEDs are increasingly being used in much larger arrays of up to 300 components as stop/slow/go lights at urban and suburban intersections.
Multi-color combinations of pluralities of visible colored LEDs are being used as the source of projected white light for illumination in binary-complementary and ternary RGB illuminators. Such illuminators are useful as vehicle or aircraft maplights, for example, or as vehicle or aircraft reading or courtesy lights, cargo lights, license plate illuminators, backup lights, and exterior mirror puddle lights. Other pertinent uses include portable flashlights and other illuminator applications where rugged, compact, lightweight, high efficiency, long-life, low voltage sources of white illumination are needed. Phosphor-enhanced xe2x80x9cwhitexe2x80x9d LEDs may also be used in some of these instances as illuminators.
Infrared (IR) emitting LEDs are being used for remote control and communication in such devices as VCR, TV, CD and other audio-visual remote control units. Similarly, high intensity IR-emitting LEDs are being used for communication between IRDA devices such as desktop, laptop and palmtop computers, PDAs (personal digital assistants), and computer peripherals such as printers, network adapters, pointing devices (xe2x80x9cmicexe2x80x9d, trackballs, etc.), keyboards and other computers. IR LED emitters and IR receivers also serve as sensors for proximity or presence in industrial control systems, for location or orientation within such opto-electronic devices such as pointing devices and optical encoders, and as read heads in such systems as barcode scanners. Blue, violet and UV emitting LEDs and LED lasers are being used extensively for data storage and retrieval applications such as reading and writing to high-density optical storage disks.
Billions of LED components are used in applications such as those cited hereinabove, in part because relatively few standardized LED configurations prevail and due to the fact that these configurations are readily processed by the automated processing equipment used almost universally by the world""s electronic assembly industries. Automated processing via mainstream equipment and procedures contributes to low capital cost, low defect rates, low labor cost, high throughput, high precision, high repeatability and flexible manufacturing practices. Without these attributes, the use of LEDs becomes cost prohibitive or otherwise unattractive from a quality standpoint for most high-volume applications.
Two of the most important steps in modern electronic assembly processes are high-speed automated insertion and mass automated soldering. Compatibility with automatic insertion or placement machines and one or more common mass soldering process are critical to large-scale commercial viability of discrete semiconductor optical emitters (including LEDs).
Thus the vast majority of LEDs used take the form of discrete-packaged THD (Through Hole Device) or SMD (Surface Mount Device) components. These configurations primarily include radial-lead THD configurations known as xe2x80x9cT-1xe2x80x9d and xe2x80x9cT-1xc2xexe2x80x9d or similar devices with rectangular shapes, all of which are readily adapted onto tape-and-reel or tape-and-ammo packaging for convenient shipment, handling and high speed automated insertion into printed circuit boards on radial inserters. Other common discrete THD LED packages include axial components such as the xe2x80x9cpolyLEDxe2x80x9d which are readily adapted onto tape and reel for convenient shipment, handling and high speed automated insertion into printed circuit boards on axial inserters. Common SMD LED components such as the xe2x80x9cTOPLEDxe2x80x9d and Pixar are similarly popular as they are readily adapted into blister-pack reels for convenient shipment, handling and high speed automated placement onto printed circuit boards with chip shooters.
Soldering is a process central to the manufacture of most conventional circuit assemblies using standardized discrete electronic devices, whether THD or SMD. By soldering the leads or contacts of a discrete electronic component such as an LED to a printed circuit board (PCB), the component becomes electrically connected to electrically conductive traces on the PCB and also to other proximal or remote electronic devices used for supplying power to, controlling or otherwise interacting electronically with the discrete electronic device. Soldering is generally accomplished by wave solder, IR reflow solder, convective IR reflow solder, vapor phase reflow solder or hand soldering. Each of these approaches differ from one another, but they all produce substantially the same end effectxe2x80x94inexpensive electrical connection of discrete electronic devices to a printed circuit board by virtue of a metallic or inter-metallic bond. Wave and reflow solder processes are known for their ability to solder a huge number of discrete devices en masse, achieving very high throughput and low cost, along with superior solder bond quality and consistency.
Widely available cost-effective alternatives to wave solder and reflow solder processes for mass production do not presently exist. Hand soldering suffers from inconsistency and high cost. Mechanical connection schemes are expensive, bulky and generally ill suited for large numbers of electrical connections in many circuits. Conductive adhesives such as silver-laden epoxies may be used to establish electrical connections on some circuit assemblies, but these materials are more costly and expensive to apply than solder. Spot soldering with lasers and other selective-solder techniques are highly specialized for specific configurations and applications and may disrupt flexible manufacturing procedures preferred in automated electronic circuit assembly operations. Thus compatibility with wave solder or reflow solder processes are de facto requirements of an effective semiconductor optical emitter component. The impact of this requirement is far reaching, because these solder operations can introduce large thermal stresses into an electronic component sufficient to degrade or destroy the component. Thus an effective semiconductor optical emitter component must be constructed in such a fashion as to protect the device""s encapsulation and encapsulated wire bonds, die attach and chip from transient heat exposure during soldering.
Conventional solder processes require that the ends of the leads of the device (below any standoff or at a point where the leads touch designated pads on the PCB) be heated to the melting point of the solder for a sustained period. This profile can include temperature excursions at the device leads as high as 230-300 degrees C. for as long as 15 seconds. Given that the leads of the device are normally constructed of plated metals or alloys such as copper or steel, this high temperature transient poses no problems for the leads themselves. The problem instead is the ability of these leads to conduct heat along their length into the encapsulated body of the device. Since these heated leads are in contact with the interior of the body of the device, they temporarily raise the local internal temperature of the device during solder processing. This can harm the somewhat delicate encapsulation, encapsulated wire bonds, die attach and chip. This phenomenon represents one of the fundamental limitations of low-cost, opto-electronic semiconductor devices today.
Keeping the body of an electronic component from rising excessively above the glass transition temperature of its encapsulating material during solder processing is critical, since the Coefficient of Thermal Expansion of polymer encapsulating materials rises dramatically above their glass transition points, typically by a factor of 2 or more. Polymers will increasingly soften, expand and plastically deform above their glass transition points. This deformation from polymer phase transition and thermal expansion in encapsulants can generate mechanical stress and cumulative fatigue severe enough to damage a discrete semiconductor device, resulting in poor performance of the device and a latent predisposition toward to premature field failure. Such damage typically consists of: 1) fatigue or fracture of electrical wire bonds (at the chip bond pads or at the lead-frame); 2) partial delamination or decomposition of die-attach adhesive; 3) micro-fracture of the chip itself, and 4) degradation of the device encapsulant, especially near the entry points of the leads into the encapsulant, and a compromised ability to seal out environmental water vapor, oxygen or other damaging agents.
With regard to such thermal vulnerability, a crucial difference must be recognized between encapsulating materials suitable for non-optical electronic devices and those suitable for optical devices. The encapsulants used for non-optical devices may be opaque, whereas those sed in constructing opto-electronic emitters and receivers must be substantially transparent in the operating wavelength band of the device. The side effects of this distinction are subtle and far ranging.
Since there is no need for transparency in non-optical devices, encapsulating materials for non-optical semiconductor devices may include a wide range of compositions containing a variety of opaque polymer binders, cross-linking agents, fillers, stabilizers and the like. Compositions of this type such as heavily filled epoxy may possess high glass transition temperatures (Tg), low thermal expansion coefficients (Cte) and/or elevated thermal conductivity such that they are suitable for transient exposures up to 175 degrees C. Opaque ceramic compositions may be thermally stable up to several hundred degrees C., with no significant phase transition temperatures to worry about, extremely low Cte and elevated thermal conductivity. For these reasons, exposure of conventional, opaque encapsulation materials for non-optical devices to electrical leads heated to 130 degrees C. or more for 10 seconds or so (by a solder wave at 230-300 degrees C.) is not normally a problem.
However, the need for optical transparency in encapsulants for opto-electronic emitters and receivers obviates use of most high-performance polymer-filler blends, ceramics and composites that are suitable for non-optical semiconductors. Without the presence of inorganic fillers, cross-linking agents or other opaque additives, the clear polymer materials used to encapsulate most opto-electronic devices are varieties of epoxies having low Tg values, greater Cte and low thermal conductivity. As such, they are not suitable for exposure to transient temperature extremes greater than about 130 degrees C.
To compensate for the potentially severe effects of damage from solder processing, prior art opto-electronic devices have undertaken a variety of improvements and compromises. The most notable improvement has been the relatively recent introduction of clear epoxies for encapsulation capable of enduring temperatures 10 to 20 degrees C. higher than those previously available (up to 130 degrees C. now vs. the previous 110 degrees C.). While useful, this has only partially alleviated the problems notedxe2x80x94the newest materials in use still fall 50 degrees C. or more short of parity with conventional non-optical semiconductor encapsulation materials.
The most common compromise used to get around the transient temperature rise problem associated with soldering is to simply increase the thermal resistance of the electrical leads used in the device construction. By increasing the thermal resistance of these solderable leads, the heat transient experienced within the device body during soldering is minimized. Such an increase in thermal resistance can typically be accomplished in the following manner without appreciably affecting the electrical performance of the leads: 1) using a lead material with lower thermal conductivity (such as steel); 2) increasing the stand-off length of the leads (distance between solder contact and the device body); or 3) decreasing the cross-sectional area of the leads.
Using these 3 techniques, prior art devices have been implemented with elevated thermal resistance of the electrical leads to provide the desired protection from the solder process.
While effective at protecting prior art devices from thermal transients associated with soldering, there are limits to this approach, particularly in the application of high power semiconductor opto-electronic emitters. Increased lead thermal resistance results in elevated internal operating temperatures in prior art devices, severely compromising operational performance and reliability of these devices. The soldered electrical leads of most prior art LED devices conduct power to the device and serve as the primary thermal dissipation path for heat created within the device during operation. Thus the electrical leads in prior art devices must be configured to possess thermal resistance as low as possible to facilitate heat extraction during normal operation. Radiation and natural convection from prior art devices play only a minor role in transferring internal heat to ambient, and thermal conduction through their encapsulating media is severely impeded by the low thermal conductivity of the optical materials used. Therefore, the electrically and thermally conductive metal leads must extract a majority of the heat to ambient by the mechanism of conduction. Greater thermal resistance in the solderable pins of these devices, necessary to protect the device from the transient thermal effects of soldering operations, therefore causes a higher internal temperature rise within the encapsulated device body during operation.
The maximum temperature rise of a portion of the device body in contact with the semiconductor emitter under steady state is approximately equal to the product of the power dissipation of the emitter and the thermal resistance between the emitter and the ambient environment.
As previously discussed, severe consequences will result if the device internal temperature rises substantially above the encapsulant Tg value. Above this temperature, the Cte of the encapsulant typically increases very rapidly, producing great thermo-mechanical stress and cumulative fatigue at the LED wirebond and attach. For most mobile applications suchs automobiles, aircraft and the like, ambient temperatures commonly reach 80 degrees C. With encapsulation maximum operating temperatures in the range of 130 degrees C., an opto-electronic emitter for these applications must therefore limit its operational xcex94T to an absolute maximum of about 50 degrees C. This in turn limits the power that can be dissipated in a given component, and in turn limits the current that may be passed through the component. Since the emitted flux of semiconductor optical emitters are typically proportional to the electrical current passed through them, limitations upon maximum electrical current also create limitations on flux generated.
Other fundamental properties of LEDs place further restrictions on the useful xcex94T for operation. Semiconductor LEDs including IR, visible and UV emitters emit light via the physical mechanism of electro-luminescence. Their emission is characteristic of the band gap of the materials from which they are composed and their quantum efficiency varies inversely with their internal temperature. An increase in LED chip temperature results in a corresponding decrease in their emission efficiency. This effect is quite significant for all common types of LEDs for visible, UV and IR emission. Commonly, a 1 degree C. increase (xcex94T) in chip temperature results in a 1% reduction in useful radiation and a 0.1 nm shift in the peak wavelength of the emission, assuming operation is at a constant power. Thus, a xcex94T of 40 degrees C. will typically result in a 40% reduction in emitted flux and a 4 nm shift in peak wavelength. Considering the effects discussed hereinabove, LED performance as a function of package thermal design may be approximated by Equation 1:
"PHgr"Ta=Ixc3x97excex2xc2x7Rxc2x7(axc2x7I2+bxc2x7I)xe2x80x83xe2x80x83(Equation 1)
Where:
"PHgr"Ta=Normalized LED Flux
xcex2=Coefficient for Change in LED emitted flux with respect to temperature (typically xe2x88x920.011)
I=the average current of the device at equilibrium
R=is the thermal resistance of the package.
a=Constant related to the equivalent series resistance of the emitter (ranging from 2.0-7.0, typically 5.5)
b=Constant related to the minimum forward voltage of the emitter (ranging typically 1.2-5.0)
Equation 1 is illustrated in FIG. 1 for various power levels and package designs.
From the preceding discussion, it can be seen that to avoid thermal damage and achieve optimal LED emission performance, it is very important to minimize the xcex94T experienced by the LED device chip and package during operation. This can only be achieved by limiting power or reducing thermal resistance.
Limiting LED power, of course, is antithetical to the purpose of high power LEDs, i.e. to produce more useful radiation. Generating higher flux with an LED generally requires higher current (and therefore higher power). Most prior art devices, however, exhibit relatively high thermal resistance from their semiconductor radiation emitter to ambient and are compelled to limit power dissipation in order to avoid internal damage. Thus the best 5 mm T-1xc2xe THD packages are limited to about 110 mW continuous power dissipation at 25 degrees C. ambient temperature.
Other prior art devices have avoided these constraints, but have achieved high performance only by ignoring the needs of standardized, automated electronic assembly operations and adopting configurations incompatible with these processes. Still other prior art devices have achieved high performance by employing unusually expensive materials, sub-components or processes in their own construction.
For example, one prior art approach that has been used to overcome these limitations uses hermetic semiconductor packaging, hybrid chip-on-board techniques, exotic materials such as ceramics, KOVAR and glass, or complex assemblies instead of or in addition to polymer encapsulation. While relevant for certain high-cost aerospace and telecommunications applications (where component cost is not a significant concern), such devices require expensive materials and unusual assembly processes. This results in high cost and restricted manufacturing capacityxe2x80x94both of which effectively preclude the use of such components in mass-market applications. The device disclosed in U.S. Pat. No. 4,267,559 by Johnson et al., illustrates a good example of this.
Johnson et al. discloses a device which includes both a TO-18 header component and a heat coupling means for mounting an LED chip thereto and transferring internally generated heat to external heat dissipating means. The header consists of several components including a KOVAR member, insulator sleeves and electrical posts, and is manufactured in a specialized process to ensure that the posts are electrically insulated as they pass through the header. The heat coupling means is a separate component from the header and is composed of copper, copper alloys, aluminum or other high thermal conductivity materials. According to the teachings of U.S. Pat. No. 4,267,559, the KOVAR header subassembly and copper heat coupling means must be bonded together with solder or electrically conductive adhesive for electrical continuity, allowing flow of electrical current into the heat coupling means and subsequently into the LED chip. Furthermore, the header and heat coupling means of U.S. Pat. No. 4,267,559 are made of completely dissimilar materials and must be so because of their unique roles in the described assembly. The header must be made of KOVAR in order that it may have a similar coefficient of thermal expansion to the insulator sleeves that run through it. At least one such sleeve is necessary to electrically isolate electrical pins from the header itself. But KOVAR has relatively low thermal conductivity, necessitating the inclusion of a separate heat coupling means made of a material such as copper with a higher thermal conductivity. Since the header is a complex subassembly itself and is made of different materials than the heat coupling means, it must be made separately from the heat coupling means and then later attached to the heat coupling means with solder or an electrically-conductive adhesive.
LED devices made similar to those described in U.S. Pat. No. 4,267,559 are currently being marketed in specialized forms similar to a TO-66 package. These devices are complex and typically involve insulated pin and header construction, and/or include specialty sub-components such as ceramic isolation sheets within them.
Another approach which has been used to avoid damage to opto-electronic emitters from soldering has been to prohibit soldering of the component altogether or to otherwise require use of laser spot soldering or other unusual electrical attachment method. This can allow construction of a device with low thermal resistance from the semiconductor emitter to the electrical pins without danger of device damage from soldering operations. The SnapLED and Super SnapLED devices made by Hewlett Packard illustrate this approach. In these devices, electrical connections are made to circuitry by mechanically stamping the leads to a simple circuit rather than soldering. The resultant devices are capable of continuous power dissipation as high as 475 mW at room temperature. This configuration, however, may complicate integration of such components with electronic circuits having higher complexityxe2x80x94such circuits are conventionally made using printed circuit boards, automated insertion equipment and wave or reflow solder operations.
A final approach is illustrated by an LED package called the SuperFlux package, available from HP. The SuperFlux device combines moderate thermal resistance between the encapsulated chip and the solder standoff on the pins with a high-grade optical encapsulant and specialized chip materials and optical design. It achieves a moderate power dissipation capability without resorting to a non-solderable configuration such as the SnapLED. However there are several significant problems with this configuration that inhibit its broader use.
The package geometry of the SuperFlux package renders it incompatible with conventional high-speed THD radial or axial insertion machinery or by SMT chip shooters known to the present inventors. Instead, it must be either hand-placed or placed by expensive, slow, robotic odd-form insertion equipment. The SuperFlux package geometry is configured for use as an xe2x80x9cend-onxe2x80x9d source onlyxe2x80x94no readily apparent convenient lead-bend technique can convert this device into a 90-degree xe2x80x9cside-lookerxe2x80x9d source. The moderate thermal resistance of the solderable pins of this device and relatively low heat capacity may leave it vulnerable to damage from poorly controlled solder processes. It may be inconvenient or costly for some electronic circuit manufacturers to control their soldering operations to the degree needed for this configuration. Finally, there is no convenient mechanism known to the inventors to outfit a SuperFlux package with a conventional active or passive heat sink.
A principle factor impeding further application of these and other LED devices in signaling, illumination and display applications is that there is not currently available a device that has a high power capability with high emitted flux where the device is easily adaptable to automated insertion and/or mass soldering processes. These limitations have either impeded the practical use of LEDs in many applications requiring high flux emission, or they have mandated the use of arrays of many LED components to achieve desired flux emission.
Consequently, it is desirable to provide a semiconductor optical emitter device that combines high emission output with thermal protection during automated processing.